Input balanced state-space realization based adaptive recursive filters

It is well known that adaptive infinite impulse response (IIR) filters can offer high performance in many applications and that the conventional direct-form parameterization yields a simple implementation but suffers from potential instability problems and slow convergence. In this paper, based on an input balanced state-space realization a new adaptive IIR filter is derived. Besides other nice properties, this realization is characterized with a parametrization under which the stability of IIR filters is ensured all the time and therefore, like the lattice parametrization based adaptive IIR filters, no stability monitoring is needed. Simulations show that the proposed adaptive IIR filters out-perform those parameterized with the direct-form realization and the normalized lattice parametrization.